The development of valid methods for assessing child behavior problems across development is critical for etiologic and prevention research.. Preschool children present to mental health clinics for services and demonstrate impaired functioning as a result of behavior problems. Yet currently there is no widely endorsed method for assessing clinically significant problem behaviors in preschool children. The Diagnostic and Statistical Manual-Fourth Edition (DSM-IV) is the most widely used nosologic system for children, but the majority of the validation sample consisted of school-age children and adolescents. Furthermore, there is debate as to whether the preschool behavioral and emotional problems should be considered within a diagnostic framework at all. Thus, there is to whether preschool behavioral and emotional problems should be considered with a diagnostic framework at all. Thus, there is an urgent need for empirical investigation that examines whether behavioral disorders can be validly diagnosed in young children within the DSM-IV framework. The proposed longitudinal study will address these issues by testing: 1) the concurrent validity of DSM-IV ODD and CD for preschool children within a developmental framework including observations of child behavior; 2) the construct validity of DSM-V ODD and CD for preschool children by examining whether risk factors established for school-age ODD and CD are associated with preschool ODD and CD; and 3) the predictive validity of DSM-IV ODD and CD by examining the stability of these disorders from preschool to school- age, and the factors that affect stability. These aims will be tested in a sample of 240 3-5 year old children from low to middle income, urban environments. Half of the participants will be clinically referred for disruptive behavior problems and half will be a comparison group matched on gender and ethnicity. If DSM-IV diagnoses of ODD and CD are valid for preschoolers, then etiologic and prevention research must focus on the first years of life.